180 ALASKAN GLACIER STUDIES
assuming its possibility, it is easily proved not to be the case here. This theory fails to account for the advance of the glacier fronts, the spreading at the sides, and the thickening of the glaciers. It fails to account for the breaking of some glaciers and the lack of breaking in others, for the breaking of the Marvine lobe of the Malaspina and the immunity of the rest of this piedmont ice area, and for the appearance of breaking in different glaciers in successive years. It cannot be seriously entertained.
The second earthquake hypothesis is that the earthquakes of 1899 broke the glaciersr but that the effects of the breaking did not appear until ablation revealed the cracks. This theory does not explain the glacier advance, spreading and thickening, and is, consequently, inapplicable here. It is therefore hardly necessary to consider whether a glacier could be broken below the surface and not on the surface, and also whether ablation could reveal the breaking in a period of about 10 months, most of which was winter season, and at the same time in a moraine-covered piedmont bulb as in the clear ice area of the valley glacier as far up as the snow line.
A third earthquake hypothesis is that the severe shaking in 1899 resulted in the setting up of unusual conditions of strain in the ice, or of disturbing equilibrium in the glacier, whose manifold points of contact with the irregular valley bottom and side are postulated to have complex strain relationships, resulting in advance of the glacier in the restoration of equilibrium. No such strain relationships are known to exist and, if they do, it is not known that they are competent to cause forward movement. The hypothesis does not account for the immunity of most of the steep-grade glaciers from advance, for these should seek equilibrium first. It accounts for no supply to keep the forward movement going and thrust a glacier front forward two miles. It may be objected to first on the ground of postulating conditions not known to exist, and, second, in crediting them with great powers where slight power, if any, is more probable.
These three earthquake theories are so entirely inapplicable that it would not be deemed worth while to give space to a consideration of them, were it not for the fact that each has been proposed by a colleague as a possible alternate hypothesis for the theory which is stated in the following section.
Theory of Earthquake Avalanche Supply. This theory is based upon three postulates, each of which is deduced from actual observation. The first postulate is that in normal conditions the mountain slopes, above a certain level, are deeply covered with snow and ice wherever it can stand, and that much of it is in such unstable position that sufficient quantities are normally sliding into the valleys to form the glaciers and keep them supplied. The second postulate is that by September, in any year, a considerable addition to the snow accumulation has been made by the early autumn snows, and that at that season the snow line has been materially lowered by the accumulation of snow, on slopes from which much of it would normally be removed by the next summer's melting. The third and vitally important postulate is that the continued shaking of the mountains during the earthquakes in the first three weeks of September, 1899, dislodged vast quantities of this snow and ice and threw it in avalanches down upon the glaciers and upon the neves. This added to the glaciers a sudden and great supply of snow and ice, some of which never would have reached the glaciers, but most of which would have come down to them from time to time during a series of years.
From these basal postulates we have inferred that so great was the sudden increase of snow and ice in the glacier reservoirs that a wave of advance was started of far greater